Diode-pumped (DP), solid-state (SS) lasers running at high pulse repetition rates are employed widely in a variety of applications including laser micromachining. In these lasers, the pulse width is largely determined by the resonator design and affected by laser pumping level and pulse repetition rate for a given laser medium. Once the resonator is constructed, there are few practical ways for changing the pulse width or temporal power profile for a given laser medium at a given pumping level and pulse repetition rate. However, for some applications such as processing links, particularly in stacks, better control of the pulse width and temporal energy profile while maintaining other laser pulse parameters is desirable.
Lasers employing a fast diode master oscillator/fiber amplifier (MOPA) configuration can deliver substantially square energy profile laser pulses with an adjustable pulse width of about 1 ns-10 ns, but current fiber amplifiers deliver random unpolarized laser output and typically have a disadvantageous wider wavelength spectrum output that imposes practical difficulties in achieving focused beam spot sizes that are sufficiently small to perform the desired micromachining operations without adversely affecting nearby substrates or other materials. U.S. patent application Ser. Nos. 10/921,481 and 10/921,765 of Sun et al., which are assigned to the assignee of this patent application, describe ways to obtain MOPA pulses with energy profiles that are specially tailored to particular applications.
Electro-optic (E-O) devices can also be employed as optical gates to reshape the energy profile or pulse width of laser pulses. However, operating the E-O devices at high repetition rates, especially above about 40 kHz, and synchronizing the laser pulses and the action(s) of an E-O device to a desired accuracy are extremely difficult to achieve under practical constraints.
Laser pulses emitted by two independent lasers that generate pulses having respectively different temporal energy profiles and pulse widths can be theoretically combined to provide a combined energy profile and pulse width of desirable features. However, in practice, combinations of such independently produced pulses suffer from laser pulse jitter, which is a random fluctuation of laser pulse initiation relative to laser pulse initiation control signals that is inherent to typical Q-switched lasers. In many applications, the pulse jitter is often greater than 5 ns-30 ns, depending on the laser design and the laser pulse repetition rate. This jitter is often too large to facilitate pulse combination with desirable accuracy, especially when the sets of combined pulses are desired to occur at intervals of less than 200 ns. For example, consistent and reproducible pulse energy profiles for an application like laser link processing could demand a timing stability between the two pulses of better than 1 ns. This synchronization problem becomes more significant at high repetition rates, especially above about 40 kHz, for example.